CA3152812A1 - Facial recognition method and apparatus - Google Patents
Facial recognition method and apparatus Download PDFInfo
- Publication number
- CA3152812A1 CA3152812A1 CA3152812A CA3152812A CA3152812A1 CA 3152812 A1 CA3152812 A1 CA 3152812A1 CA 3152812 A CA3152812 A CA 3152812A CA 3152812 A CA3152812 A CA 3152812A CA 3152812 A1 CA3152812 A1 CA 3152812A1
- Authority
- CA
- Canada
- Prior art keywords
- image
- recognized
- face
- compared
- features
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 230000001815 facial effect Effects 0.000 title claims abstract description 65
- 238000000034 method Methods 0.000 title claims abstract description 49
- 230000036544 posture Effects 0.000 claims abstract description 14
- 230000008569 process Effects 0.000 claims description 18
- 238000001514 detection method Methods 0.000 claims description 16
- 238000010606 normalization Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 3
- 230000014509 gene expression Effects 0.000 abstract description 4
- 241000282414 Homo sapiens Species 0.000 description 6
- 210000000887 face Anatomy 0.000 description 6
- 230000008859 change Effects 0.000 description 5
- 238000012937 correction Methods 0.000 description 4
- 238000005286 illumination Methods 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 210000003128 head Anatomy 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 210000000554 iris Anatomy 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 210000000056 organ Anatomy 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/172—Classification, e.g. identification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/168—Feature extraction; Face representation
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Collating Specific Patterns (AREA)
Abstract
Disclosed are a facial recognition method and apparatus. The method comprises: acquiring a recognition request and an image to be recognized; according to the image to be recognized, acquiring a facial feature of the image to be recognized; according to the recognition request, acquiring, from a comparison library, all features of a reference image for comparison, wherein each ID in the comparison library comprises reference images of different postures in a plurality of different scenarios; and according to the facial feature of the image to be recognized, and all the features of the reference image for comparison, determining whether the image to be recognized matches the reference image for comparison. According to the present invention, by using the comparison library of reference images, of the same ID, of different postures in a plurality of different scenarios as a matching standard, the recognition accuracy is improved, the robustness of an algorithm is enhanced, and a better adaptability to the expression, image quality, etc. of an image to be recognized is provided.
Description
FACIAL RECOGNITION METHOD AND APPARATUS
BACKGROUND OF THE INVENTION
Technical Field [0001] The present invention relates to the field of computer vision technology, and more particularly to a face recognizing method and a face recognizing device.
Description of Related Art
BACKGROUND OF THE INVENTION
Technical Field [0001] The present invention relates to the field of computer vision technology, and more particularly to a face recognizing method and a face recognizing device.
Description of Related Art
[0002] Face recognition is a biometric technology that performs identification based on facial feature information of human beings. The series of relevant techniques of using video cameras or cameras to collect images or video streams containing human faces, and automatically detecting and tracking human faces in the images to recognize the detected faces are usually referred to as poi ______________________________________ tiait recognition, and face recognition. With the development of the technology, the face recognition technique has become one of the hottest applications in artificial intelligence, such as boarding by face scanning, extracting toilet tissue by face scanning, paying by face scanning, clocking in and out by face scanning, and recognizing pedestrians running the red light by face scanning, etc.
[0003] There are usually three application modes for face recognition:
1. The 1-to-1 mode is also referred to as identification verification mode, it is essentially a process whereby a computer rapidly compares a current face with a poi _______________________________________________________________________ tiait database to see whether there is a match, and this can be simply understood as verifying that you are you. Boarding, ticket checking and paying by "face swiping"
all pertain to the 1-to-1 verification of person with certificate.
2. The 1-to-N mode is to find out face data of a current user (namely image to be Date Recue/Date Received 2022-02-28 recognized) from a portrait database with colossal volume and to match.
Cracking down on abduction and exposing persons who run the red light all pertain to the 1-to-N face recognition, namely to find out 1 target from N number of faces.
3. The M-to-N mode is a process whereby a computer performs face recognition on all persons within a scenario and compares the same with a porn _______________ ait database, and this process is a dynamic face comparison fully applicable to a plurality of scenarios, such as public security, guests welcoming, robot application, etc.
1. The 1-to-1 mode is also referred to as identification verification mode, it is essentially a process whereby a computer rapidly compares a current face with a poi _______________________________________________________________________ tiait database to see whether there is a match, and this can be simply understood as verifying that you are you. Boarding, ticket checking and paying by "face swiping"
all pertain to the 1-to-1 verification of person with certificate.
2. The 1-to-N mode is to find out face data of a current user (namely image to be Date Recue/Date Received 2022-02-28 recognized) from a portrait database with colossal volume and to match.
Cracking down on abduction and exposing persons who run the red light all pertain to the 1-to-N face recognition, namely to find out 1 target from N number of faces.
3. The M-to-N mode is a process whereby a computer performs face recognition on all persons within a scenario and compares the same with a porn _______________ ait database, and this process is a dynamic face comparison fully applicable to a plurality of scenarios, such as public security, guests welcoming, robot application, etc.
[0004] However, due to such factors as the increase in the number of comparisons in the 1-to-1 mode and the increase in the number of registered underlying library IDs in the 1-to-N
mode, the probabilities of matching failure and matching error apparently increase, and this is so mainly because facial features change significantly. Factors affecting precision of face recognition include the following:
1) Facial features are not fixedly unchanged, as they apparently change with ages, expressions, gestures and making-up;
2) Sheltering by ornaments, goggles, mouth masks or other objects renders faces incomplete and also increases the difficulty for the recognition;
3) Such collecting environments of training sets as illumination and equipment parameters restrict the performance of algorithm models on testing sets of other scenarios;
4) Similarities between different persons might be increased due to similar making-up mode and modeling, and wearing of similar ornaments by the different persons.
mode, the probabilities of matching failure and matching error apparently increase, and this is so mainly because facial features change significantly. Factors affecting precision of face recognition include the following:
1) Facial features are not fixedly unchanged, as they apparently change with ages, expressions, gestures and making-up;
2) Sheltering by ornaments, goggles, mouth masks or other objects renders faces incomplete and also increases the difficulty for the recognition;
3) Such collecting environments of training sets as illumination and equipment parameters restrict the performance of algorithm models on testing sets of other scenarios;
4) Similarities between different persons might be increased due to similar making-up mode and modeling, and wearing of similar ornaments by the different persons.
[0005] Accordingly, there is an urgent need to propose a novel face recognizing method to overcome the aforementioned many problems.
SUMMARY OF THE INVENTION
Date Recue/Date Received 2022-02-28
SUMMARY OF THE INVENTION
Date Recue/Date Received 2022-02-28
[0006] In order to solve the problems pending in the state of the art, embodiments of the present invention provide a face recognizing method and a face recognizing device, so as to overcome such prior-art problems as low precision of face recognition due to such factors as the increase in the number of comparisons and the increase in the number of registered underlying library IDs.
[0007] In order to solve one or more of the aforementioned technical problem(s), the present invention employs the following technical solutions.
[0008] According to one aspect, there is provided a face recognizing method, and the method comprises the following steps:
[0009] obtaining a recognition request and an image to be recognized;
[0010] obtaining a facial feature of the image to be recognized according to the image to be recognized;
[0011] obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios; and
[0012] judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
[0013] Further, the step of obtaining a facial feature of the image to be recognized according to the image to be recognized includes:
[0014] performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image to which the image to be recognized corresponds and key-point positions;
[0015] performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and
[0016] performing feature extraction on the processed face image, to obtain a facial feature to Date Recue/Date Received 2022-02-28 which the image to be recognized corresponds.
[0017] Further, when the recognition request is a 1-to-1 recognition request, the recognition request includes an ID of the reference image to be compared; and
[0018] the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
[0019] obtaining all features of all reference images under the ID of the reference image to be compared from the comparison library according to the ID of the reference image to be compared.
[0020] Further, the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes:
[0021] comparing all features of each reference image under the ID of the reference image to be compared with the facial feature of the image to be recognized, and determining that the reference image is successfully matched with the image to be recognized if similarity exceeds a first preset threshold; and
[0022] obtaining the number of the reference images successfully matched with the image to be recognized, and determining that the image to be recognized is successfully matched with the ID of the reference image to be compared if the number exceeds half of the total number of the reference images.
[0023] Further, when the recognition request is a 1-to-N recognition request, the reference images to be compared are all reference images under all IDs in the comparison library;
and
and
[0024] the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
[0025] obtaining all features of all reference images under all IDs from the comparison library according to the recognition request.
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
[0026] Further, the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes:
[0027] calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID and the facial feature of the image to be recognized; and
[0028] judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
[0029] According to another aspect, there is provided a face recognizing device that comprises:
[0030] a data obtaining module, for obtaining a recognition request and an image to be recognized;
[0031] a first feature obtaining module, for obtaining a facial feature of the image to be recognized according to the image to be recognized;
[0032] a second feature obtaining module, for obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios; and
[0033] an image recognizing module, for judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
[0034] Further, the first feature obtaining module includes:
[0035] an image detecting unit, for performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image and key-point positions to which the image to be recognized corresponds;
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
[0036] a normalization processing unit, for performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and
[0037] a feature extracting unit, for performing feature extraction on the processed face image, to obtain a facial feature to which the image to be recognized corresponds.
[0038] Further, when the recognition request is a 1-to-1 recognition request, the recognition request includes an ID of the reference image to be compared;
[0039] the second feature obtaining module is specifically employed for:
[0040] obtaining all features of all reference images under the ID of the reference image to be compared from the comparison library according to the ID of the reference image to be compared.
[0041] Further, the image recognizing module includes:
[0042] a first calculating unit, for calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID of the reference image to be compared and the facial feature of the image to be recognized;
[0043] a first comparing unit, for comparing the similarity with a first preset threshold, and determining that the reference image is successfully matched with the image to be recognized if the similarity is greater than the first preset threshold; and
[0044] a second comparing unit, for obtaining the number of the reference images successfully matched with the image to be recognized, and determining that the image to be recognized is successfully matched with the ID of the reference image to be compared if the number exceeds half of the total number of the reference images.
[0045] Further, when the recognition request is a 1-to-N recognition request, the reference images to be compared are all reference images under all IDs in the comparison library;
[0046] the second feature obtaining module is further employed for:
[0047] obtaining all features of all reference images under all IDs from the comparison library according to the recognition request.
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
[0048] Further, the image recognizing module further includes:
[0049] a second calculating unit, for calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID
and the facial feature of the image to be recognized; and
and the facial feature of the image to be recognized; and
[0050] a third comparing unit, for judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
[0051] The technical solutions provided by the embodiments of the present invention bring about the following advantageous effects:
[0052] By using reference images of different postures under different scenarios of the same ID
in a comparison library as the matching standard, the face recognizing method and face recognizing device provided by the embodiments of the present invention enhance precision of recognition, strengthen robustness of the algorithm, and exhibit excellent adaptability to the expression and picture quality of the image to be recognized.
in a comparison library as the matching standard, the face recognizing method and face recognizing device provided by the embodiments of the present invention enhance precision of recognition, strengthen robustness of the algorithm, and exhibit excellent adaptability to the expression and picture quality of the image to be recognized.
[0053] By employing the solution of detection together with face frame and face key points, the face recognizing method and face recognizing device provided by the embodiments of the present invention can not only precisely locate face position, but can also reduce steps and time in the recognizing process, and enhance recognition efficiency.
[0054] By firstly calculating the similarity between each reference image under each ID and the image to be recognized under the 1-to-N recognition scenario, the face recognizing method and face recognizing device provided by the embodiments of the present invention select the ID with the highest similarity, then judge whether the similarity satisfies a second preset threshold to thereby match out the ID matched with the image to Date Recue/Date Received 2022-02-28 be recognized, whereby anti-attack capability of the algorithm is enhanced.
BRIEF DESCRIPTION OF THE DRAWINGS
BRIEF DESCRIPTION OF THE DRAWINGS
[0055] To more clearly describe the technical solutions in the embodiments of the present invention, drawings required to illustrate the embodiments will be briefly introduced below. Apparently, the drawings introduced below are merely directed to some embodiments of the present invention, while persons ordinarily skilled in the art may further acquire other drawings on the basis of these drawings without spending creative effort in the process.
[0056] Fig. 1 is a flowchart illustrating the face recognizing method in an exemplary embodiment;
[0057] Fig. 2 is a flowchart illustrating obtaining a facial feature of the image to be recognized according to the image to be recognized in an exemplary embodiment;
[0058] Fig. 3 is a flowchart illustrating judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared in the 1-to-1 mode in an exemplary embodiment;
[0059] Fig. 4 is a flowchart illustrating judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared in the 1-to-N
mode in an exemplary embodiment; and
mode in an exemplary embodiment; and
[0060] Fig. 5 is a view schematically illustrating the structure of the face recognizing device in an exemplary embodiment.
DETAILED DESCRIPTION OF THE INVENTION
DETAILED DESCRIPTION OF THE INVENTION
[0061] To make more lucid and clear the objectives, technical solutions and advantages of the present invention, the technical solutions in the embodiments of the present invention will be clearly and comprehensively described below with reference to the accompanying Date Recue/Date Received 2022-02-28 drawings in the embodiments of the present invention. Apparently, the embodiments as described are merely partial, rather than the entire, embodiments of the present invention.
Any other embodiments makeable by persons ordinarily skilled in the art on the basis of the embodiments in the present invention without creative effort shall all fall within the protection scope of the present invention.
Any other embodiments makeable by persons ordinarily skilled in the art on the basis of the embodiments in the present invention without creative effort shall all fall within the protection scope of the present invention.
[0062] Facial features are a type of biometric features most suitable for calibrating identifications, and are advantageous in terms of simple collection, low cost and easy recognition relative to such features as fingerprints and irises. The use of human faces to perform identifications has been widely applied in many such scenarios as login through face swiping, credit investigation by face swiping, and security verification, etc.
Face recognition on the one hand reduces manual operation and economizes on production cost, and on the other hand facilitates identification verification, and enhances user experience. The basic process of face recognition is to extract features from a face image to be recognized, and thereafter compare the features with facial features already registered in a database.
Face recognition on the one hand reduces manual operation and economizes on production cost, and on the other hand facilitates identification verification, and enhances user experience. The basic process of face recognition is to extract features from a face image to be recognized, and thereafter compare the features with facial features already registered in a database.
[0063] Fig. 1 is a flowchart illustrating the face recognizing method in an exemplary embodiment.
With reference to Fig. 1, the method comprises the following steps.
With reference to Fig. 1, the method comprises the following steps.
[0064] Si - obtaining a recognition request and an image to be recognized.
[0065] Specifically, the image to be recognized is collected from a picture, a video or a camera, and a corresponding recognition request is obtained. Face recognition usually includes three modes, namely 1-to-1 mode, 1-to-N mode, and M: N mode. As should be noted, the face recognizing method provided by the embodiments of the present invention is mainly applicable to the 1-to-1 recognition scenario and the 1-to-N recognition scenario, so the recognition request in the embodiments of the present invention mainly includes a 1-to-1 recognition request or a 1-to-N recognition request. 1-to-1 matching mainly solves the Date Recue/Date Received 2022-02-28 problem of determining whether the image to be recognized and the reference image pertain to the same person, whereas 1-to-N matching mainly solves the problem of determining to which person the image to be recognized pertains.
[0066] S2 - obtaining a facial feature of the image to be recognized according to the image to be recognized.
[0067] Specifically, before face recognition is performed on the image to be recognized, it is firstly required to extract the facial feature of the image to be recognized from the image to be recognized, as a preferred example, it is possible in the embodiments of the present invention to extract the facial feature of the image to be recognized (namely a depth feature of the image to be recognized) by means of convolutional neural network.
[0068] S3 - obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios.
[0069] Specifically, the recognition request in the embodiments of the present invention mainly includes a 1-to-1 recognition request or a 1-to-N recognition request. While recognition is being specifically performed, the mode is employed to compare the facial feature of the image to be recognized with features of the reference image to be compared. Since reference images to be compared as employed by different recognition requests are also different, it is therefore further required to obtain all features of the corresponding reference image to be compared from a comparison library according to the recognition request.
[0070] Each ID in the comparison library contains reference images of different postures under different scenarios, so as to ensure diversity of the comparison library and to enhance the probability of successful recognition. In the embodiments of the present invention, the Date Recue/Date Received 2022-02-28 reference images maintained under each ID in the comparison library mainly subsume the following three scenarios:
[0071] Scenario A: ID photos authorized to be usable for personal identifications
[0072] ID photos uploaded by enterprise or corporate members and authorized for use as personal identifications of internal personnel can be used as reference images in a scenario.
[0073] Scenario B: Authorized, registered photos of logins through face scanning
[0074] The function of login through face scanning of the socializing software seeks authorization from users to use their head poi ____________________________ hafts for personal identifications. During the cooperative registration process of the software, users are required to perform such actions as nodding and shaking their heads, while cameras check these actions, automatically collect face images turned properly leftwards and rightwards and store the same under corresponding IDs to serve as reference images.
[0075] Scenario C: Photos collected for internal security
[0076] In an internal office site, the terminal equipment for security monitoring accumulates a certain amount of face pictures for security reasons. Out of the consideration for security, these photos can also serve as reference images in a scenario. When photos collected for security are stored under corresponding IDs to serve as reference images, these photos are firstly clustered before they are matched with the comparison library, and the specific process is as follows.
[0077] Firstly, the picture quality and orientation of the camera that collects photos for security are checked and proofread, to ensure the collection of photos with suitable angles and illumination. Qualified photos for security are subsequently registered in a temporary underlying library; if similarity between the latest collected photo for security and the most matched registered photo in the underlying library exceeds a threshold, the collected photo is put under the corresponding ID, otherwise a new ID is registered therefor.
Whenever a photo is newly added to the ID, the registered photo is updated once, the updated registered photo should have the highest similarity with other photos under the Date Recue/Date Received 2022-02-28 ID, and this registered photo is considered to be most characteristic of the ID. The registered photo as obtained is taken as a photo to be recognized, and the most similar ID
is matched by the logic of 1 to N out of the comparison library to serve as the ID of the photo for security.
Whenever a photo is newly added to the ID, the registered photo is updated once, the updated registered photo should have the highest similarity with other photos under the Date Recue/Date Received 2022-02-28 ID, and this registered photo is considered to be most characteristic of the ID. The registered photo as obtained is taken as a photo to be recognized, and the most similar ID
is matched by the logic of 1 to N out of the comparison library to serve as the ID of the photo for security.
[0078] As should be additionally noted, during collection of reference images, the following quality screening standards should be satisfied: multi-dimensional thresholds such as definition, angles and action amplitudes of the facial organs should be satisfied, and the reference images as screened out should take the cases of frontal face and slightly turned face both into consideration, so as to enhance the probability of successfully matching with images to be recognized of plural postures.
[0079] As should be further noted, if the reference images in the comparison library are collected through a 1-to-1 port, when they are being registered in the comparison library, a video of the registering person is recorded, a quality algorithm is employed to obtain several photos with the best quality from the video for feature extraction and registration in the library. If collection is performed through a 1-to-N port, after the photo to be recognized has been collected, it is compared with existing reference images in the comparison library, if similarity of the most similar reference image is also lower than the preset threshold, it is considered that there is no reference image of the corresponding ID in the underlying library, in which case the person in the image to be recognized is tracked, a video is collected therefor, the quality algorithm is employed to select several photos with the best quality from the video, and a new ID is created and added to the registration library.
[0080] S4 - judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
[0081] Specifically, the similarity between the image to be recognized and the reference image to be compared is calculated according to the facial feature of the image to be recognized and all the features of the reference image to be compared, and it is then judged according to the similarity whether the image to be recognized is matched with the reference image to be compared.
[0082] Fig. 2 is a flowchart illustrating obtaining a facial feature of the image to be recognized according to the image to be recognized in an exemplary embodiment. With reference to Fig. 2, as a preferred mode of execution in the embodiments of the present invention, the step of obtaining a facial feature of the image to be recognized according to the image to be recognized includes the following steps:
[0083] S2.1 - performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image and key-point positions to which the image to be recognized corresponds.
[0084] Specifically, performing face frame detection on the image to be recognized means to detect and locate the human face in the image to be recognized, to return a high-precision face frame coordinate, and to cut out the face image from the image to be recognized according to the coordinate. Performing face key-point detection on the image to be recognized is to locate such key region positions of the human face as the eyes, nose and mouth, etc. There may be one or more face image(s) detected from the image to be recognized, and the user can base on the actual application scenario to decide whether to recognize a single face frame or plural face frames, to which no restriction is made in the embodiments of the present invention.
[0085] S2.2 - performing a normalization process on the face image according to the key-point positions, to obtain a processed face image.
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
[0086] Specifically, performing a normalization process on the face image aims to provide consistency to photos of the same person captured under different imaging conditions (such as illumination intensities, directions, distances, and postures, etc.), so as to facilitate subsequent extraction of facial features. Face normalization process includes the contents of two aspects, one is geometric normalization, the other is grayscale normalization. Also referred to as position correction, geometric normalization helps correct size difference and angle inclination caused by change in imaging distances and face postures, and can solve the problems concerning face scale change and face rotation.
Specifically included are the three links of face scale normalization, planar face rotation correction (head tilting), and deep face rotation correction (face twisting).
A 3D model of face can be utilized for some deep face rotation corrections with higher requirements.
Grayscale normalization is employed to compensate for face images obtained under different light intensities and light source directions, so as to weaken changes in image signals caused purely by change in illumination. As should be noted here, to facilitate subsequent use of a model (such as convolutional neural network) to extract facial features, it is further required in the embodiments of the present invention to adjust the face image to a size suitable for input into the model.
Specifically included are the three links of face scale normalization, planar face rotation correction (head tilting), and deep face rotation correction (face twisting).
A 3D model of face can be utilized for some deep face rotation corrections with higher requirements.
Grayscale normalization is employed to compensate for face images obtained under different light intensities and light source directions, so as to weaken changes in image signals caused purely by change in illumination. As should be noted here, to facilitate subsequent use of a model (such as convolutional neural network) to extract facial features, it is further required in the embodiments of the present invention to adjust the face image to a size suitable for input into the model.
[0087] S2.3 - performing feature extraction on the processed face image, to obtain a facial feature to which the image to be recognized corresponds.
[0088] Specifically, by way of example, a pre-trained convolutional neural network is employed in the embodiments of the present invention to perform feature extraction on the face image processed by the aforementioned steps to obtain a facial feature to which the image to be recognized corresponds.
[0089] As a preferred mode of execution in the embodiments of the present invention, when the recognition request is a I-to-1 recognition request, the recognition request includes an ID
of the reference image to be compared; and Date Recue/Date Received 2022-02-28
of the reference image to be compared; and Date Recue/Date Received 2022-02-28
[0090] the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
[0091] obtaining all features of all reference images under the ID of the reference image to be compared from the comparison library according to the ID of the reference image to be compared.
[0092] Specifically, when the face recognition mode is a 1-to-1 mode, then the recognition request as obtained is a 1-to-1 recognition request, in which case it is required to recognize whether the image to be recognized and the reference image pertain to the same person, so the recognition request includes the ID of the reference image to be compared. At this time, all features of all reference images under the ID are obtained from the comparison library according to the ID of the reference image to be compared.
[0093] Fig. 3 is a flowchart illustrating judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared in an exemplary embodiment. With reference to Fig. 3, as a preferred mode of execution in the embodiments of the present invention, the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes the following steps:
[0094] S401 - calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID of the reference image to be compared and the facial feature of the image to be recognized.
[0095] Specifically, when the face recognition mode is a 1-to-1 mode, it is required to compare each reference image under the ID of the reference image to be compared with the image to be recognized. During specific implementation, it is possible to calculate the similarity Date Recue/Date Received 2022-02-28 between each reference image and the image to be recognized according to all features of each reference image under the ID of the reference image to be compared and the facial feature of the image to be recognized, and to compare the images according to the similarity.
[0096] S402 - comparing the similarity with a first preset threshold, and determining that the reference image is successfully matched with the image to be recognized if the similarity is greater than the first preset threshold.
[0097] Specifically, the first preset threshold can be set as practically required, and the specific numerical value of the first preset threshold is not restricted in this context. When the similarity between the reference image and the image to be recognized exceeds (is greater than) the first preset threshold, it is determined that the reference image is successfully matched with the image to be recognized.
[0098] S403 - obtaining the number of the reference images successfully matched with the image to be recognized, and determining that the image to be recognized is successfully matched with the ID of the reference image to be compared if the number exceeds half of the total number of the reference images.
[0099] Specifically, when the number of the reference images under the ID of the reference image to be compared successfully matched with the image to be recognized exceeds half of the total number (which means the total number of the reference images under the ID
of the reference image to be compared) of the reference images participating in the matching, it is finally determined that the image to be recognized is successfully matched with the ID of the reference image to be compared, otherwise it is determined that the image to be recognized is not successfully matched with the ID of the reference image to be compared. Such setting makes it possible to avoid lower similarity between reference images of the same ID and a single image to be recognized as caused by relatively large Date Recue/Date Received 2022-02-28 differences in scenarios and shelterings, etc., and it is also made possible to remove such problems as unduly high similarity between reference images of different IDs and the image to be recognized as caused by accidental factors, so as to enhance recognition precision.
of the reference image to be compared) of the reference images participating in the matching, it is finally determined that the image to be recognized is successfully matched with the ID of the reference image to be compared, otherwise it is determined that the image to be recognized is not successfully matched with the ID of the reference image to be compared. Such setting makes it possible to avoid lower similarity between reference images of the same ID and a single image to be recognized as caused by relatively large Date Recue/Date Received 2022-02-28 differences in scenarios and shelterings, etc., and it is also made possible to remove such problems as unduly high similarity between reference images of different IDs and the image to be recognized as caused by accidental factors, so as to enhance recognition precision.
[0100] As a preferred mode of execution in the embodiments of the present invention, when the recognition request is a 1-to-N recognition request, the reference images to be compared are all reference images under all IDs in the comparison library; and
[0101] the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
[0102] obtaining all features of all reference images under all IDs from the comparison library according to the recognition request.
[0103] Specifically, when the face recognition mode is a 1-to-N mode, then the recognition request as obtained is a 1-to-N recognition request, in which case it is required to recognize to which person the image to be recognized pertains, so the reference images to be compared are all reference images under all IDs in the comparison library. At this time, it is required to obtain all features of all reference images under all IDs from the comparison library.
[0104] Fig. 4 is a flowchart illustrating judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared in the 1-to-N
mode in an exemplary embodiment. With reference to Fig. 4, as a preferred mode of execution in the embodiments of the present invention, the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes the following steps.
Date Recue/Date Received 2022-02-28
mode in an exemplary embodiment. With reference to Fig. 4, as a preferred mode of execution in the embodiments of the present invention, the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes the following steps.
Date Recue/Date Received 2022-02-28
[0105] S501 - calculating to obtain similarity between the image to be recognized and each ID
according to all features of each reference image under each ID and the facial feature of the image to be recognized.
according to all features of each reference image under each ID and the facial feature of the image to be recognized.
[0106] Specifically, when the face recognition mode is a 1-to-N mode, it is required to compare all reference images under all IDs in the comparison library with the image to be recognized. During specific implementation, it is possible to calculate the similarity between each reference image and the image to be recognized according to all features of each reference image under each ID and the facial feature of the image to be recognized, and to compare the images according to the similarity.
[0107] S502 ¨judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
[0108] Specifically, all reference images can be firstly sequenced according to the order of the similarity values, the ID to which the reference image with the highest similarity with the image to be recognized corresponds is selected, it is subsequently judged whether the similarity value satisfies a second preset threshold, if yes, it is determined that the image to be recognized is matched with the ID with the highest similarity value, if not, it is determined that the image to be recognized is an unregistered image. As should be noted here, under the circumstance in which each ID has a plurality of reference images under a plurality of scenarios, the probability for the image to be recognized to be matched with a reference image under its own ID as the maximal similarity is enhanced, and the problem is avoided in which similarity is low between a single image to be recognized and the reference images under its own ID due to accidental factors. Thusly, the second preset threshold can be set as practically required, the specific numerical value of the second preset threshold is not restricted in this context, and the second preset threshold is Date Recue/Date Received 2022-02-28 set in order to shield photos with unregistered IDs against attack.
[0109] As a preferred mode of execution in the embodiments of the present invention, the image to be recognized that is not successfully matched with the reference image to be compared is registered in the comparison library, a new ID is generated, and the image to be recognized is stored to the new ID to serve as a reference image under the new ID.
[0110] Fig. 5 is a view schematically illustrating the structure of the face recognizing device in an exemplary embodiment. With reference to Fig. 5, the device comprises:
[0111] a data obtaining module, for obtaining a recognition request and an image to be recognized;
[0112] a first feature obtaining module, for obtaining a facial feature of the image to be recognized according to the image to be recognized;
[0113] a second feature obtaining module, for obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios; and
[0114] an image recognizing module, for judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
[0115] As a preferred mode of execution in the embodiments of the present invention, the first feature obtaining module includes:
[0116] an image detecting unit, for performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image and key-point positions to which the image to be recognized corresponds;
[0117] a normalization processing unit, for performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and
[0118] a feature extracting unit, for performing feature extraction on the processed face image, Date Recue/Date Received 2022-02-28 to obtain a facial feature to which the image to be recognized corresponds.
[0119] As a preferred mode of execution in the embodiments of the present invention, when the recognition request is a 1-to-1 recognition request, the recognition request includes an ID
of the reference image to be compared;
of the reference image to be compared;
[0120] the second feature obtaining module is specifically employed for:
[0121] obtaining all features of all reference images under the ID of the reference image to be compared from the comparison library according to the ID of the reference image to be compared.
[0122] As a preferred mode of execution in the embodiments of the present invention, the image recognizing module includes:
[0123] a first calculating unit, for calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID of the reference image to be compared and the facial feature of the image to be recognized;
[0124] a first comparing unit, for comparing the similarity with a first preset threshold, and determining that the reference image is successfully matched with the image to be recognized if the similarity is greater than the first preset threshold; and
[0125] a second comparing unit, for obtaining the number of the reference images successfully matched with the image to be recognized, and determining that the image to be recognized is successfully matched with the ID of the reference image to be compared if the number exceeds half of the total number of the reference images.
[0126] As a preferred mode of execution in the embodiments of the present invention, when the recognition request is a 1-to-N recognition request, the reference images to be compared are all reference images under all IDs in the comparison library;
[0127] the second feature obtaining module is further employed for:
[0128] obtaining all features of all reference images under all IDs from the comparison library according to the recognition request.
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
[0129] As a preferred mode of execution in the embodiments of the present invention, the image recognizing module further includes:
[0130] a second calculating unit, for calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID
and the facial feature of the image to be recognized; and
and the facial feature of the image to be recognized; and
[0131] a third comparing unit, for judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
[0132] To sum it up, the technical solutions provided by the embodiments of the present invention bring about the following advantageous effects:
1. By using reference images of different postures under different scenarios of the same ID in a comparison library as the matching standard, the face recognizing method and face recognizing device provided by the embodiments of the present invention enhance precision of recognition, strengthen robustness of the algorithm, and exhibit excellent adaptability to the expression and picture quality of the image to be recognized.
2. By employing the solution of detection together with face frame and face key points, the face recognizing method and face recognizing device provided by the embodiments of the present invention can not only precisely locate face position, but can also reduce steps and time in the recognizing process, and enhance recognition efficiency.
3. By firstly calculating the similarity between each reference image under each ID and the image to be recognized under the 1-to-N recognition scenario, the face Date Recue/Date Received 2022-02-28 recognizing method and face recognizing device provided by the embodiments of the present invention select the ID with the highest similarity, then judge whether the similarity satisfies a second preset threshold to thereby match out the ID
matched with the image to be recognized, whereby anti-attack capability of the algorithm is enhanced.
1. By using reference images of different postures under different scenarios of the same ID in a comparison library as the matching standard, the face recognizing method and face recognizing device provided by the embodiments of the present invention enhance precision of recognition, strengthen robustness of the algorithm, and exhibit excellent adaptability to the expression and picture quality of the image to be recognized.
2. By employing the solution of detection together with face frame and face key points, the face recognizing method and face recognizing device provided by the embodiments of the present invention can not only precisely locate face position, but can also reduce steps and time in the recognizing process, and enhance recognition efficiency.
3. By firstly calculating the similarity between each reference image under each ID and the image to be recognized under the 1-to-N recognition scenario, the face Date Recue/Date Received 2022-02-28 recognizing method and face recognizing device provided by the embodiments of the present invention select the ID with the highest similarity, then judge whether the similarity satisfies a second preset threshold to thereby match out the ID
matched with the image to be recognized, whereby anti-attack capability of the algorithm is enhanced.
[0133] As should be noted, when the face recognizing device provided by the aforementioned embodiment triggers a face recognizing business, it is merely exemplarily described with its division into the aforementioned various functional modules, whereas in actual application it is possible to base on requirements to assign the aforementioned functions to different functional modules for completion, that is to say, the internal structure of the device is divided into different functional modules to complete the entire or partial functions as described above. In addition, the face recognizing device provided by the aforementioned embodiment pertains to the same inventive conception as the face recognizing method, in other words, the device is based on the face recognizing method ¨ see the method embodiment for its specific implementation process, while no repetition will be made in this context.
[0134] As comprehensible to persons ordinarily skilled in the art, the entire or partial steps in the aforementioned embodiments can be completed via hardware, or via a program instructing relevant hardware, the program can be stored in a computer-readable storage medium, and the storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.
[0135] The foregoing embodiments are merely preferred embodiments of the present invention, and they are not to be construed as restrictive to the present invention. Any amendment, equivalent substitution, and improvement makeable within the spirit and principle of the present invention shall all fall within the protection scope of the present invention.
Date Recue/Date Received 2022-02-28
Date Recue/Date Received 2022-02-28
Claims (10)
1. A face recognizing method, characterized in comprising the following steps:
obtaining a recognition request and an image to be recognized;
obtaining a facial feature of the image to be recognized according to the image to be recognized;
obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios; and judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
obtaining a recognition request and an image to be recognized;
obtaining a facial feature of the image to be recognized according to the image to be recognized;
obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios; and judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
2. The face recognizing method according to Claim 1, characterized in that the step of obtaining a facial feature of the image to be recognized according to the image to be recognized includes:
performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image to which the image to be recognized corresponds and key-point positions;
performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and performing feature extraction on the processed face image, to obtain a facial feature to which the image to be recognized corresponds.
performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image to which the image to be recognized corresponds and key-point positions;
performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and performing feature extraction on the processed face image, to obtain a facial feature to which the image to be recognized corresponds.
3. The face recognizing method according to Claim 1 or 2, characterized in that, when the recognition request is a 1-to-1 recognition request, the recognition request includes an ID of the reference image to be compared;
the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
obtaining all features of all reference images under the ID of the reference image to be compared from the comparison library according to the ID of the reference image to be compared.
the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
obtaining all features of all reference images under the ID of the reference image to be compared from the comparison library according to the ID of the reference image to be compared.
4. The face recognizing method according to Claim 3, characterized in that the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes:
calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID of the reference image to be compared and the facial feature of the image to be recognized;
comparing the similarity with a first preset threshold, if the similarity is greater than the first preset threshold, determining that the reference image is successfully matched with the image to be recognized; and obtaining the number of the reference images successfully matched with the image to be recognized, if the number exceeds half of the total number of the reference images, determining that the image to be recognized is successfully matched with the ID of the reference image to be compared.
calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID of the reference image to be compared and the facial feature of the image to be recognized;
comparing the similarity with a first preset threshold, if the similarity is greater than the first preset threshold, determining that the reference image is successfully matched with the image to be recognized; and obtaining the number of the reference images successfully matched with the image to be recognized, if the number exceeds half of the total number of the reference images, determining that the image to be recognized is successfully matched with the ID of the reference image to be compared.
5. The face recognizing method according to Claim 1 or 2, characterized in that, when the recognition request is a 1-to-N recognition request, the reference images to be compared are all reference images under all IDs in the comparison library;
the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
obtaining all features of all reference images under all IDs from the comparison library according to the recognition request.
the step of obtaining all features of a reference image to be compared from a comparison library according to the recognition request includes:
obtaining all features of all reference images under all IDs from the comparison library according to the recognition request.
6. The face recognizing method according to Claim 5, characterized in that the step of judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared includes:
calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID and the facial feature of the image to be recognized; and judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID and the facial feature of the image to be recognized; and judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
7. A face recognizing device based on the method according to any of Claims 1 to 6, characterized in that the device comprises:
a data obtaining module, for obtaining a recognition request and an image to be recognized;
a first feature obtaining module, for obtaining a facial feature of the image to be recognized according to the image to be recognized;
a second feature obtaining module, for obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios;
and an image recognizing module, for judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
a data obtaining module, for obtaining a recognition request and an image to be recognized;
a first feature obtaining module, for obtaining a facial feature of the image to be recognized according to the image to be recognized;
a second feature obtaining module, for obtaining all features of a reference image to be compared from a comparison library according to the recognition request, wherein each ID in the comparison library contains reference images of different postures under different scenarios;
and an image recognizing module, for judging whether the image to be recognized is matched with the reference image to be compared according to the facial feature of the image to be recognized and all the features of the reference image to be compared.
8. The face recognizing device according to Claim 7, characterized in that the first feature obtaining module includes:
an image detecting unit, for performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image to which the image to be recognized corresponds and key-point positions;
a normalization processing unit, for performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and a feature extracting unit, for performing feature extraction on the processed face image, to obtain a facial feature to which the image to be recognized corresponds.
an image detecting unit, for performing face frame detection and face key-point detection on the image to be recognized, to obtain a face image to which the image to be recognized corresponds and key-point positions;
a normalization processing unit, for performing a normalization process on the face image according to the key-point positions, to obtain a processed face image; and a feature extracting unit, for performing feature extraction on the processed face image, to obtain a facial feature to which the image to be recognized corresponds.
9. The face recognizing device according to Claim 7 or 8, characterized in that the image recognizing module includes:
a first calculating unit, for calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID
of the reference image to be compared and the facial feature of the image to be recognized;
a first comparing unit, for comparing the similarity with a first preset threshold, and determining that the reference image is successfully matched with the image to be recognized if the similarity is greater than the first preset threshold; and a second comparing unit, for obtaining the number of the reference images successfully matched with the image to be recognized, and determining that the image to be recognized is successfully matched with the ID of the reference image to be compared if the number exceeds half of the total number of the reference images.
a first calculating unit, for calculating similarity between each reference image and the image to be recognized according to all features of each reference image under the ID
of the reference image to be compared and the facial feature of the image to be recognized;
a first comparing unit, for comparing the similarity with a first preset threshold, and determining that the reference image is successfully matched with the image to be recognized if the similarity is greater than the first preset threshold; and a second comparing unit, for obtaining the number of the reference images successfully matched with the image to be recognized, and determining that the image to be recognized is successfully matched with the ID of the reference image to be compared if the number exceeds half of the total number of the reference images.
10. The face recognizing device according to Claim 7 or 8, characterized in that the image recognizing module further includes:
a second calculating unit, for calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID
and the facial feature of the image to be recognized; and a third comparing unit, for judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
a second calculating unit, for calculating to obtain similarity between the image to be recognized and each ID according to all features of each reference image under each ID
and the facial feature of the image to be recognized; and a third comparing unit, for judging whether similarity of the ID with the highest similarity value satisfies a second preset threshold, if yes, determining that the image to be recognized is matched with the ID with the highest similarity value, if not, determining that the image to be recognized is an unregistered image.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910793283.2A CN110688901A (en) | 2019-08-26 | 2019-08-26 | Face recognition method and device |
| CN201910793283.2 | 2019-08-26 | ||
| PCT/CN2020/096992 WO2021036436A1 (en) | 2019-08-26 | 2020-06-19 | Facial recognition method and apparatus |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CA3152812A1 true CA3152812A1 (en) | 2021-03-04 |
Family
ID=69108460
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA3152812A Pending CA3152812A1 (en) | 2019-08-26 | 2020-06-19 | Facial recognition method and apparatus |
Country Status (3)
| Country | Link |
|---|---|
| CN (1) | CN110688901A (en) |
| CA (1) | CA3152812A1 (en) |
| WO (1) | WO2021036436A1 (en) |
Families Citing this family (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110688901A (en) * | 2019-08-26 | 2020-01-14 | 苏宁云计算有限公司 | Face recognition method and device |
| CN111368721B (en) * | 2020-03-03 | 2023-05-05 | 深圳市腾讯计算机系统有限公司 | Identity recognition processing method and device, electronic equipment and storage medium |
| CN111694979A (en) * | 2020-06-11 | 2020-09-22 | 重庆中科云从科技有限公司 | Archive management method, system, equipment and medium based on image |
| CN112084903A (en) * | 2020-08-26 | 2020-12-15 | 武汉普利商用机器有限公司 | Method and system for updating face recognition base photo |
| CN112016508B (en) * | 2020-09-07 | 2023-08-29 | 杭州海康威视数字技术股份有限公司 | Face recognition method, device, system, computing device and storage medium |
| CN112667840B (en) * | 2020-12-22 | 2024-05-28 | 中国银联股份有限公司 | Feature sample library construction method, traffic identification method, device and storage medium |
| CN113033476B (en) * | 2021-04-19 | 2022-08-12 | 清华大学 | Cross-posture face recognition method |
| CN113542348B (en) * | 2021-05-27 | 2022-09-06 | 武汉旷视金智科技有限公司 | Image data transmission method and device |
| CN113569676B (en) * | 2021-07-16 | 2024-06-11 | 北京市商汤科技开发有限公司 | Image processing method, device, electronic equipment and storage medium |
| CN113362324B (en) * | 2021-07-21 | 2023-02-24 | 上海脊合医疗科技有限公司 | Bone health detection method and system based on video image |
| CN113536270B (en) * | 2021-07-26 | 2023-08-08 | 网易（杭州）网络有限公司 | Information verification method, device, computer equipment and storage medium |
| CN113688764A (en) * | 2021-08-31 | 2021-11-23 | 瓴盛科技有限公司 | Training method and device for face optimization model and computer readable medium |
| CN115880761B (en) * | 2023-02-09 | 2023-05-05 | 数据空间研究院 | Face recognition method, system, storage medium and application based on policy optimization |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101770613A (en) * | 2010-01-19 | 2010-07-07 | 北京智慧眼科技发展有限公司 | Social insurance identity authentication method based on face recognition and living body detection |
| CN102004908B (en) * | 2010-11-30 | 2012-10-17 | 汉王科技股份有限公司 | Self-adapting face identification method and device |
| US8873813B2 (en) * | 2012-09-17 | 2014-10-28 | Z Advanced Computing, Inc. | Application of Z-webs and Z-factors to analytics, search engine, learning, recognition, natural language, and other utilities |
| CN102542299B (en) * | 2011-12-07 | 2015-03-25 | 惠州Tcl移动通信有限公司 | Face recognition method, device and mobile terminal capable of recognizing face |
| GB2499449A (en) * | 2012-02-20 | 2013-08-21 | Taiwan Colour And Imaging Technology Corp | Surveillance by face recognition using colour display of images |
| CN104463237B (en) * | 2014-12-18 | 2018-03-06 | 中科创达软件股份有限公司 | A kind of face verification method and device based on multi-pose identification |
| CN110688901A (en) * | 2019-08-26 | 2020-01-14 | 苏宁云计算有限公司 | Face recognition method and device |
-
2019
- 2019-08-26 CN CN201910793283.2A patent/CN110688901A/en active Pending
-
2020
- 2020-06-19 CA CA3152812A patent/CA3152812A1/en active Pending
- 2020-06-19 WO PCT/CN2020/096992 patent/WO2021036436A1/en active Application Filing
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021036436A1 (en) | 2021-03-04 |
| CN110688901A (en) | 2020-01-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA3152812A1 (en) | Facial recognition method and apparatus | |
| US11188734B2 (en) | Systems and methods for performing fingerprint based user authentication using imagery captured using mobile devices | |
| KR102483642B1 (en) | Method and apparatus for liveness test | |
| US20170262472A1 (en) | Systems and methods for recognition of faces e.g. from mobile-device-generated images of faces | |
| JP2020501264A (en) | System and method for performing fingerprint authentication using images captured with a mobile device | |
| US10769415B1 (en) | Detection of identity changes during facial recognition enrollment process | |
| US10922399B2 (en) | Authentication verification using soft biometric traits | |
| US11651624B2 (en) | Iris authentication device, iris authentication method, and recording medium | |
| US11610437B2 (en) | Face liveness detection systems and methods | |
| CN110008943B (en) | Image processing method and device, computing equipment and storage medium | |
| US20220044014A1 (en) | Iris authentication device, iris authentication method and recording medium | |
| KR101718244B1 (en) | Apparatus and method of processing wide angle image for recognizing face | |
| US20230161854A1 (en) | Fool-proof registration system and methods thereof | |
| CN116189316A (en) | Living body detection method and system | |
| CN114926717A (en) | Object feature fusion method and device and storage medium | |
| CN116266419A (en) | Living body detection method and device and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request |
Effective date: 20220916 |
|
| EEER | Examination request |
Effective date: 20220916 |
|
| EEER | Examination request |
Effective date: 20220916 |
|
| EEER | Examination request |
Effective date: 20220916 |